The Namespace Management Of Distribute File System

With the rapid development of internet, more and more materials need to be stored inthe disk. The amount of global information grow more than30percent over a year, and thedata that needs to be stored to the disk grow more than114percent over a year. Traditionalfile system like ext2, ext3and net file system (NFS) can’t feed our needs, At this moment,distribute file system come into our sight. Distribute file system can store1PB to dozensof PB data, it can feed us well. Traditional file system typically uses3party architecture,and nameserver is the brain of all nodes. Nameserver typically faces C10K problem, sohow to design the nameserver to adapt it to high concurrency problem becomes more andmore important.This paper design and implement the distribute file system named RaccoonFS. Thepaper also analyzes and solves the problem that nameserver behaves bad after highconcurrency because of the lock, it proposes a new namespace management method.RacoonFS uses B+tree to manage namespace metadata, and implement copy on write andmultiversion concurrency control on the B+tree, so the read operations and writeoperations separate from each other and write operations can do concurrently. The thingsabove contribute much to the performance of RaccoonFS.In the test part, we choose three kinds of object to test: the hadoop distribute filesystem, the B+tree namespace management that has implemented copy on write, the B+tree namespace management that has implemented copy on write and multiversionconcurrency control. The result show that the B+tree namespace management that hasimplemented copy on write and multiversion concurrency control behaves better than theB+tree namespace management that has implement copy on write, and the B+treenamespace management that has implement copy on write behaves better than hadoopdistribute file system. The test also shows that copy on write and multiversion concurrencycontrol can contribute much to namespace management.